80% of People Analytics Is Not Spent on Analysis

If you take a university course on people analytics, you will probably learn a lot about statistics and nothing at all about automation. That’s a pity because in the real world of people analytics, you will spend very little time using statistics — whereas automation can be a big part of your life.

By automation, I mean using the power of the computer to ease your workload by taking over some tasks, even just some small ones. Automation might mean some carefully programmed Robotic Process Automation that handles the whole process of a standard set of analyses. It might also mean a simple Excel macro that manipulates a table of data. Or it could even be just the use of a built-in automation tool like a pivot table that makes certain types of analysis much easier.

The reason automation is so important in analytics is that, typically, more of an analyst’s time is spent cleaning and organizing data rather than analyzing it. The work of an analyst involves a whole series of small tasks to pull data out of various systems, sort it, remove outliers, and so on. I’ve had people analytics pros say it takes 70% to 80% of their time wrangling the data before they can get around to analyzing it.

One of the most common time-wasters in the life of an analytics pro is preparing data for standard reports or dashboards. Leaders may think that the analytics team is just pushing a button to get the report, but in reality, there is a tedious process of pulling data from various HR systems, organizing it all in Excel, and then perhaps moving the data into PowerPoint to become part of a standard presentation.

When we face up to the unpleasant reality that the job of a people analytics pro is more about data wrangling than data analysis, it suddenly becomes clear that expertise in automating as many steps as possible in that wrangling phase is a key skill set.

Among newcomers, there will be an eagerness to automate the whole process of any routine analytics work so they really can do it at the push of a button. That’s an admirable goal; however, it’s often more effort than it’s worth.

Seemingly routine tasks often have a surprising amount of variation, which upsets the automated process. One can end up in a situation where an automated process takes just seconds to generate an analysis, but then you need to spend hours figuring out why the result doesn’t seem plausible and uncover where the automation went awry. There is great power in automating little parts of a process with human oversight at each step of the way to keep things on track.

The takeaway here is for analytics pros to fully embrace the automation side of their work and recognize that their skills and wisdom in deploying a range of tools to automate things is a key part of the value they add. Indeed, people analytics pros may want to consider that their ability to automate work is one of their superpowers — and they should work hard to develop it.